But I’m not qualified to be a data scientist!

How to transition from university to the data science industry

Erika Braithwaite

2022-03-09

Welcome!

So you’re thinking about leaving the comforting womb of university, and transitioning to a data science career in industry? You may be feeling….

Overwhelmed by the possibilities
Confused about the terminology
Insecure about your capacities
And so many more emotions

Objectives

I’ll do my best to address these objectives in the most data-driven approach I can!

Dispel some myths about the data science industry
Tips on how to transition
Let’s talk about your CV, the job search and the interview
New directions in the data science industry
Bestow some glimmer of hope ⭐

A bit about me

I’m the CEO of a health data science startup in Montreal www.precision-analytics.ca

My background

BA Psychology, Concordia
MSc Psychology, Université de Montréal
PhD Epidemiology and Biostatistics, McGill

Our company

Founded in 2017
Began as a statistical consultancy, evolved to a blend of Software as a Service
Work in the biopharmaceutical sector to create tools that combine data storage, analysis and communication

What is a data scientist

A combination of programming, statistics and domain knowledge.

I like to think of it as story telling using data

A very scary (and real) job posting

Coding knowledge and experience with several languages: C, C++, Java, JavaScript, etc.
Knowledge and experience in statistical and data mining techniques: GLM/Regression, Random Forest, Boosting, Trees, text mining, social network analysis, etc.
Experience querying databases and using statistical computer languages: R, Python, SLQ, etc.
Experience using web services: Redshift, S3, Spark, Digital Ocean, etc.
Experience creating and using advanced machine learning algorithms and statistics: regression, simulation, scenario analysis, modeling, clustering, decision trees, neural networks, etc.
Experience analyzing data from 3rd party providers: Google Analytics, Site Catalyst, Coremetrics, Adwords, Crimson Hexagon, Facebook Insights, etc.
Experience with distributed data/computing tools: Map/Reduce, Hadoop, Hive, Spark, Gurobi, MySQL, etc.
Experience visualizing/presenting data for stakeholders using: Periscope, Business Objects, D3, ggplot, etc.

Myths about data science

Myth 1. You need a PhD to become a data scientist

Myth 2. ###

Myth 3. Data science == AI & machine learning

Myth 4. Data science is all about technical skills

Letting the data speak!

To tackle these myths, Kaggle conducts an annual survey of data scientists. Over 25,000 respondents completed the survey in 2021

‼️ Kaggle is a platform that hosts machine learning competitions. It’s users are not representative of the entire data science community ‼️

Data available for 2021

Demographics of respondents

Types of data scientist positions

Myth 1. You need a PhD to become a data scientist

Experience of respondents

Myth 2. Primary tools used at work or school

Myth 3. Data science == Machine learning and AI

Myth 4. Data science challenges are mostly analytical

The survey did not explicitly ask the way respondents broke up their time. So I did the next best thing… ask twitter

Data scientists challenges

In 2017, the Kaggle survey asked respondents about the “biggest challenges in data science”.

From here we can see that most people identify “dirty data” the toughest part of the job. The rest of the issues seem be organizational.

On transitioning

Understand your new “audience”

💰 The currency in academia versus the currency in the private sector 💰

Gaining more experience: Courses

The American Statistical Association held a two-day Data Science summit. 72 educators, researchers and practitioners in statistics, mathematics, computer science, and data science from academia, industry, government and nonprofit gathered to put forth recommendations for future data science programs.

Data science courses should expose students to:

Introduction to statistics
Data analysis in the real world
Math and algorithms
Answering real problems
Expose students to modern tools
Teach data ethics
Active learning

I’d add to this list: study design. Knowing how data was generated/collected will always inform the types of questions you can ask, and appropriate method of analysis.

Build a portfolio

Showcase your work

Github
a simple personal website
Kaggle competitions

Local meetups

RLadies, Data For Good, PyData

Internship

Get relevant experience, learn about new industry, create a network
Check out your university coop programs
Internship programs in industry IVADO labs, Plotly

Don’t forget to reach out to your network and tell people you’re looking to get into a new industry

Preparing your cover letter and CV

Use the cover letter to explain why you’re applying for the job even if your profile doesn’t perfectly match the job description
Instead of focusing on all the highlights of your academic career, explain WHY you want to work for THAT company

In general, keep it short and sweet (max 2 pages)
Academic CV != Professional CV
Only provide extensive details about academic experience if it is relevant
Tailor your CV and cover letter to each application - use key words from posting

Searching for a position

Focus on the skills listed in the job positing, not the title but be weary of descriptions
Don’t expect that you will possess 100% of the stated skills; it’s often a wishlist
Try to talk to people (ideally former employees) to get insider perspective

Be aware that many companies are jumping on the “data” bandwagon without having any infrastructure to support a junior hire

During the interview

Expect technical questions and/or coding challenge
Be honest about your skills
Be mindful about expressing your desired to be paid to learn
Highlight the value you bring to the employer, rather than the value you accrued in university
Be ready to show that you are able to explain technical concepts to non-technical audiences

Jacqueline Nolis’s book Build a Career in Data Science

“I want to hear about a project they’ve worked on recently. I ask them about how the project started, how they determined it was worth time and effort, their process, and their results. I also ask them about what they learned from the project. I gain a lot from answers to this question: if they can tell a narrative, how the problem related to the bigger picture, and how they tackled the hard work of doing something”.

Asking for feedback after your interview can help highlight the interviewer’s perceptions of your strengths and weaknesses

Employer fears

In general, employers are concerned that new hires without job experience will struggle to see the bigger picture and the pace.

Try to address these fears by addressing them directly in your cover letters, CV’s and during the interview

Conclusion

Pivoting to data science from humanities, social sciences or any other disciplines doesn’t mean you need to leave all of your training behind.

☑️ You have knowledge and/or experience in at least one of these circles

☑️ Look for people in your substantive area who are doing some more quantitative/tech driven work

☑️ Be confident that you know how to learn

✔️ Be ready to start at the bottom

☑️ Be ready be creative in your search and to be a hustler

✔️ Your path is the right path

New directions (my hot takes)

Large push towards “automating” analytics is set to change the role of data scientists (e.g. Data Robot, Alteryx, Dataiku, Google Cloud AI); storytelling is one of the last aspects of data and communication that cannot be productized

Many organizations want to innovate but are not sure how; many ML investments have failed to deliver value

In the healthcare sector, billions invested in AI for drug discovery but not a single drug has gone to market

A big thanks!

I can be found on twitter @ea_braithwaite

Come check out our website www.precision-analytics.ca

This presentation & R code can be found online https://github.com/precision-analytics/CSCDS.

Simply download the entire repo via “clone/download” button; the presentation is the html file